In radiation therapy (RT), it is essential to deliver a prescribed high radiation dose to a target volume containing malignancy while sparing surrounding normal tissues. This is accomplished by extensive use of imaging throughout the RT process. Computed tomography (CT) is the dominant imaging tool in image- guided radiation therapy (IGRT). Simulator units with cone-beam CT (CBCT) imaging capabilities have become available as part of RT planning systems. Recently, a KV X-ray imager, referred to as the on-board imager, capable of CBCT imaging has also been developed on the linear accelerator (LINAC) treatment system. The on-board imager offers a unique opportunity to yield accurate image representation of the patient before, during, and after treatment sessions. There are two broad classes of RT tasks to which these systems are suited, and each task class has its own set of requirements on image quality, acquisition speed, and patient dose minimization. The first class includes tasks such as those involving CBCT for treatment planning, where the objective is to obtain diagnostic image quality and to identify unknown object in a complex scene. The second class includes tasks such as those using CBCT for localization, where the objective is to recognize differences in position relative to the planning scan and to identify pose of known objects. The goal of the project is to, using the on-board imager as the test-bed platform, capitalize fully on the hardware capabilities and on recent algorithm advances through developing innovative scanning configurations and algorithms to yield accurate and dose-efficient volumetric images in IGRT. The specific aims of the project are: (1) To develop innovative scanning configurations for radiotherapy CBCT imaging;(2) To develop image-reconstruction algorithms for radiotherapy CBCT imaging;(3) To compensate for the physical factors in radiotherapy CBCT imaging;and (4) To evaluate the scanning configurations and algorithms for radiotherapy CBCT imaging. We have recently made significant breakthroughs in CBCT algorithms. Algorithms can be designed for accurate image reconstruction within regions-of-interest (ROI) from CBCT data acquired with X-ray illumination that partially covers the patient and with general scanning configurations. Our strategy for targeted ROI imaging resembles that of the intensity-modulated radiation therapy. It can reduce patient dose and scatter and avoid repeated illumination of critical organs. We believe that our expertise and insights developed and accumulated in our studies in both imaging and radiation therapy have placed us in a unique and strong position to perform and accomplish the proposed research on optimization of CBCT imaging in IGRT successfully and in a timely manner.